09-07-2025
Launched: Early-Stage Hodgkin Lymphoma Prediction Tool
The first prediction model to be developed for early-stage classic Hodgkin lymphoma, focusing on just four key risk factors that are assessed via an easy-to-use online risk calculator, provides individualized prediction of patients' 2-year progression-free survival, using continuous variables that enable high precision in the risk assessments.
'Utilizing objective, continuous, and readily available variables in nearly 5400 early-stage classic Hodgkin lymphoma patients, we developed and validated a robust, dynamic, and modern prediction model,' co-author Andrew M. Evens, DO, of the Division of Blood Disorders, Rutgers Cancer Institute, New Brunswick, New Jersey, said in presenting the findings at the 18th International Conference on Malignant Lymphoma (ICML) 2025 in Lugano, Switzerland.
'Male sex, and continuous values of decreasing hemoglobin or albumin, and incrementally increasing maximum tumor diameter were associated with worse progression-free survival,' Evens said of the research, which was simultaneously published in NEJM Evidence .
While the vast majority of patients with early-stage classic Hodgkin lymphoma do have favorable outcomes, the cure rate isn't 100%, and the prognostic models currently used, such as those from the European Organisation for Research and Treatment of Cancer (EORTC) or German Hodgkin Study Group (GHSG), have been used for decades, with general assessments as being either favorable or unfavorable, Evens explained.
'These original models were based on data from the early 1970s, when the majority of patients had staging laparotomies and radiation alone as treatment,' he explained. An update from the GHSG in 2013 was performed, but 'overall, it had poor specificity,' he said.
'With more sophistication available in modeling and in contemporary datasets, there has been an unmet need [for a modernized prediction tool] identified,' Evens noted.
In response, the new model was developed by Evens and colleagues as part of the HoLISTIC (Hodgkin Lymphoma International Study for Individual Care) international consortium.
Called the Early-Stage cHL International Prognostication Index (E-HIPI), the new model was developed with the use of data on 3000 patients with untreated, early-stage classic Hodgkin lymphoma from four seminal, phase 3 clinical trials, including the CCTG-ECOG, EORTC H9, RAPID, and EORTC H10 trials.
Those patients, overall, had a median age of 31.2 years, and 77.4% had stage II disease.
Their estimated 2-year progression-free survival was 93.7%.
Four Key Risk Factors in Early-Stage Classic Hodgkin Lymphoma
Based on the analysis, four key parameters that emerged as being significantly associated with progression-free survival were female sex, conveying a lower risk (hazard ratio [HR], 0.59); maximum tumor diameter (HR, 1.06 per 1 cm, ranging from 1.5 to 15.0, hence 'about a 10% increased risk with each centimeter increase,' Evens said); hemoglobin level (HR, 1.09; continuous, ranging from 5.0 to 16.5 g/dL); and albumin level (HR, 0.83; continuous, ranging from 2.5 to 6.0 g/dL), where increases from low levels to high levels were predictive factors for each.
The data were validated externally using two separate cohorts of 2360 contemporaneously treated patients from five international cancer registries, including the BC Cancer, Princess Margaret, Iowa/Mayo SPORE, Stanford Registry, and Danish National Lymphoma Registry.
In those external validation cohorts, the 2-year progression-free survival was slightly lower, at 90.3% in cohort 1 and 91.6% in cohort 2.
After multivariate adjustment in a Cox regression model, the E-HIPI model was significantly associated with progression-free survival, whereas the EORTC measures of favorable or unfavorable status were not.
The number of nodal groups was also considered as a potential predictor but was ultimately not found to be significantly associated with progression-free survival in the model.
Continuous Variables Provide Greater Context
The online risk calculator, in addition to providing a 2-year progression-free survival estimate, also helps to estimate risk according to adjustments based on differing potential disease trajectories.
Notably, whereas many other models use basic cutoffs for factors, such as age being categorized as older or younger than 45 years, the model uses continuous variables to provide context for each input in relation to a full range, instead of just being under or over a specific level.
'We know that if you try to dichotomize a continuous factor, you lose a lot of statistical power, and you lose potential nonlinear effects,' Evens explained.
With dichotomized values such as age categorized as older or younger than 45 years, 'how can you know, for instance, that outcomes with a 44-year-old are going to be similar to the 18-year-old?' he said.
'Likewise, with tumor diameters, as opposed to saying simply above or below 10 cm, this approach gives us more richness and more power in individualized prediction to base the risk on the exact tumor dimension.'
The 2-year progression-free survival was seen as the most important primary outcome for the model because among the very small proportion of patients who do have a relapse, such events most commonly occur within 2 years, Evens noted.
However, with ever-advancing therapies continuously affecting outcomes, and the known small risk for postacute late effects occurring years later, potentially due not to the disease itself but to exposure to treatments such as chemotherapy and radiation, the work on E-HIPI will continue.
The current predictor was step 1, Evens told Medscape Medical News .
Step 2 will be to look at different treatments and help predict outcomes based on the differing treatments, and step 3 will involve the estimation of postacute late effects, he explained.
'Our model provides more precise and individualized prediction [than existing methods], and in the near future with the second and third iterations of the model, we'll be able to take this to the bedside and help predict not just general outcomes for patients, but more exact treatment options,' he said.
The E-HIPI is in fact the second prognostic tool developed by the international consortium. The team has also made available the Advanced-stage cHL International Prognostication Index.
Commenting on the study, Alex Herrera, MD, chief of the Division of Lymphoma, Department of Hematology & Hematopoietic Cell Transplantation, City of Hope Medical Center in Duarte, California, said the new model offers some important new insights.
'We rely heavily on early-stage prognostic indices, since patients with early-stage cHL are classified as 'favorable' or 'unfavorable' based on these risk factor indices, and that is what determines the treatment plan,' he told Medscape Medical News .
'Here we see that some of the traditional factors in the GHSG and EORTC risk stratification may not be as useful.'
Herrera agreed that a key attribute of the new prediction model is the use of the continuous values.
'An important change [from previous models] is the use of the full range of values (continuous) as opposed to just binary thresholds,' Herrera said.
He added that 'albumin and hemoglobin have always been a part of the advanced-stage IPI, but here they were key prognostic factors for early-stage disease.'
While the previous indices will continue to determine how to treat patients, 'the online tool will allow this to become a key part of prognostic discussions with patients in the clinic,' Herrera said.